Abstract
A growing number of advanced scientific and technical applications require significant computing power both to develop results and to display those results in comprehensible form. The most effective approach for such applications has been to perform the necessary computations on supercomputers, maniframes, or minisupercomputers, and then to transport the outputs to a graphics workstation that can display the results, statically or dynamically, as three-dimensional graphics. However, current hardware and software technologies have enabled development of a new class of single-user workstation that provide these capabilities in a single platform.
This paper provides an overview of architecture of one such machine, the Titan-III.
Zusammenfassung
Eine wachsende Zahl von fortschrittlichen wissenschaftlichen und technischen Anwendungen erfordern hohe Rechenleistung sowohl zur Berechnung als auch zur verständlichen Darstellung der Ergebnisse. Die effektivste Methode für solche Anwendungen bestand darin, die nötigen Berechnungen auf Supercomputern, Mainframes oder Mini-Supercomputern durchzuführen und anschließend die Resultate zu Graphic Workstations zu liefern, die diese dann statisch oder dynamisch als 3D-Graphiken anzeigten.
Die derzeitige Hardware- und Software-Technologie hat die Entwicklung einer neuen Klasse von Single-user Workstations ermöglicht, die diese Fähigkeiten auf einer Ebene ermöglichen.
Dieser Bericht liefert einen Überblick der Architektur einer solchen Maschine, der Titan-III.
Similar content being viewed by others
References
[Alle82] Allen, J. R., Kennedy, K.: PFC: A program to convert FORTRAN to parallel form. MASC Technical Report 82-6, Rice University, 1982.
[Ande67] Anderson, D. W., Sparacio, F. J., Tomasulo, F. M.: The IBM system/360 model 91: Machine philosophy and instruction handling. IBM Journal11 (1967).
[Bell89] Bell, G., Miranker, G.: Using livermore loops and the harmonic mean to measure the basic performance of two graphics supercomputers, Vector register. October 1989.
[Buch88] Bucher, I. Y.: Vector performance of Register-to-Register computers. LA-11198-MS, Los Alamos National Laboratories, January 1988.
[Cens78] Censier, L. M., Paul Feautrier: A new solution to the coherence problems in multicache systems. IEEE Transact. Comput.C-27 (12) (1978).
[Cheu86] Cheung, T., Smith, J. E.: A simulation study of the CRAY X-MP memory system. IEEE Transact. Comput.C-35 (7) (1986).
[Deid88] Deide, T., Hagenmaier, Carl F., Miranker, G. S., Rubinstein, J., Worley, W.: The Titan graphics supercomputer. IEEE Computer, September 1988.
[Dong85] Dongarra, J. J., Duff, I. S.: Advanced architecture computers. LA-11198-MS, Los Alamos National Laboratories, January 1984.
[Hock81] Hockney, R. W., Jessope, C. R.: Parallel computers: Architecture, programming and algorithms. Bristol, England: Adams Hilger Ltd. 1981.
[Lube88] Lubeck, O. M.: Supercomputer performance: The theory, practice, and results. LA-11204-MS, Los Alamos National Laboratories, January 1988.
[Mira87] Miranker, G., Rubinstein, J., Sanguinetti, J.: Squeezing cray-class supercomputer performance into a single-user package. IEEE CompCon Proceedings, February 1987.
[Mira89] Miranker, G., Rubinstein, J., Sanguinetti, J.: Evolution in the application of ASICs in the second generation Titan. Proceedings of ICCAD '89, September 1989.
[Nico88] Nicoud, J.: Video RAMS: Structure and applications. Micro Magazine8, (1) (1988).
[Rowe86] Rowen, C., Crudele L., et al.: RISC VLSI design for system level Performance. VLSI Systems Design7 (3) (1986).
[Sufa79] Abu-Sufah, W., Kuck, D., Lawrie, D.: Automatic program translation for virtual memory computers. Proceedings 1979 National Computer Conference, June 1979.
[Thor70] Thorton, J. E.: Design of a computer — The control data 6600. Glenview, Ill.: Scott, Foresman and Co., 1970.
[Torb87] Torborg, J. W.: A parallel processor architecture for graphics arithmetic operations. Computer Graphics21 (4), 197–204 (1984) (Proc. SIGGRAPH).
Author information
Authors and Affiliations
Additional information
Dedicated to Professor W. L. Miranker on the occasion of his 60th birthday
Rights and permissions
About this article
Cite this article
Miranker, G.S. Titan III supercomputer architectural overview. Computing 48, 39–59 (1992). https://doi.org/10.1007/BF02241705
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF02241705